Authors: Michael P. Bishop*, Texas A & M University, Brennan W. Young, Texas A & M University, Jeffrey D. Colby, Appalachian State University
Topics: Remote Sensing, Earth Science
Keywords: Anisotropic Reflectance Correction, Remote Sensing, Topographic Complexity
Session Type: Paper
Start / End Time: 3:05 PM / 4:45 PM
Room: Buchanan, Marriott, Mezzanine Level
Presentation File: No File Uploaded
Assessing mountain environments using multispectral and hyperspectral imagery is notoriously difficult due to multi-scale topographic effects that govern the radiation-transfer cascade and anisotropic reflectance. A multitude of anisotropic-reflectance-correction (ARC) methods have been proposed, although researchers report conflicting results due to varying experimental designs and a lack of diagnostic evaluation procedures. Given that little progress has occurred in addressing the ARC problem in the last 40 years, we focus on addressing the fundamental principles associated with anisotropic-reflectance correction. Specifically, our research objectives are to characterize topographic complexity in an attempt to explain how topographic complexity governs our ability to address the problem and to explain conflicting results. We accomplish this by characterizing topographic complexity from a process-structure systems perspective, where we account for the dominant topographic effects governing radiation transfer and surface irradiance variations. Our surface irradiance model accounts for orbital and solar geometry variations, atmospheric conditions, and the direct and diffuse-skylight irradiance fluxes, such that we account for spatio-temporal dynamics. We compare annual surface irradiance variations and various radiation-transfer components from different mountain regions around the world to demonstrate the dynamics of topographic complexity that govern the environmental information content in imagery, and therefore, the ability of different anisotropic-reflectance correction procedures to adequately remove or reduce spectral variation caused by topography. We report on the reasons why various ARC assumptions made by the remote sensing community are not valid for interpreting the results of ARC that are based upon the coupling influence of topographic complexity and sensor resolution characteristics.